Opened 16 years ago
Closed 14 years ago
#67 closed enhancement (fixed)
new tablespace parameter for shp2pgsql
Reported by: | kneufeld | Owned by: | kneufeld |
---|---|---|---|
Priority: | medium | Milestone: | PostGIS 2.0.0 |
Component: | postgis | Version: | master |
Keywords: | Cc: |
Description (last modified by )
I would like to see the addition of a tablespace parameter added to shp2pgsql.
For loading large tables, this is currently a three stage process.
- load DDL from shp2pgsql
- use psql and alter the tablespace of the table
- load the data from shp2pgsql
Postgres stores relations as individual files. When loading several thousand tables, the underlying OS can become slow when looking up a particular file/table. Having multiple tablespaces solves this problem.
Change History (6)
comment:1 by , 16 years ago
comment:2 by , 15 years ago
Description: | modified (diff) |
---|---|
Milestone: | → postgis 1.4.1 |
Version: | → trunk |
comment:3 by , 15 years ago
Wouldn't this be considered an API change in which case it should go to 1.5
comment:4 by , 15 years ago
+1 for saving until 1.5. Otherwise we fall back into the same trap of having feature changes in point release.
ATB,
Mark.
comment:6 by , 14 years ago
Resolution: | accepted → fixed |
---|---|
Status: | assigned → closed |
Implemented on trunk, rev 6905.
You specify the tablespace for the table and the indexes (primary key and spatial index if -I is specified) separately:
-T <table tablespace>
and/or
-X <index tablespace>
I guess that could be a useful feature; as a temporary work-around you could always use ALTER USER foo SET default_tablespace 'bar' and then load the shp2pgsql output as the foo user before reverting if required.
ATB,
Mark.